Best AI apps for kids: a parent's comparison

The two questions that actually separate these products

Almost every app in this category is safe enough and polished enough to survive a demo. What separates them at primary-school age is narrower than the marketing suggests, and it comes down to two questions.

First: does your child produce something, or collect something? A tool that answers instantly and keeps no track of whether your child understood is a vending machine, however good its content is. A tool that makes them direct it, check it, and explain it back builds the habit you actually want. Second: can you see what happened without reading a transcript? Almost every product in this category has an answer to the first question in its marketing. Very few have any answer at all to the second — of the seven competing products below, two describe something a parent can actually look at, one addresses parents directly and then offers them nothing to look at, and four say nothing on the subject at all.

Both questions are answerable in one evening per product, and the checklist further down is how. The tables first.

How to tell a real kid tool from a repackaged chatbot · The four checks that tell you what a session actually was

Eight AI and creative-learning products for kids, in their own words

Every cell is quoted or paraphrased directly from the product's own live site, captured 9 October 2026. Prices and age ranges change; the capture date is on every claim for that reason, and the source links are at the foot of this page.

What the table shows once you read down the columns

Three things stand out, and none of them is a ranking.

The price spread is nearly twenty-fold, and it is not a quality spread. Scratch and Code.org are free and are two of the best things in children's computing; Khanmigo is $4 a month from a nonprofit with a very large content library behind it. If your budget is the binding constraint, you are not short of excellent options, and anyone telling you otherwise — us included — should be read with that in mind.

The stated age ranges barely overlap with your child. “4-15” and “5 to 18” are not age ranges, they are market sizes. A product built to work for a four-year-old and a fifteen-year-old at the same time is making a real trade-off somewhere, and in this category the usual place it lands is reading level: too simple at eleven, too wordy at eight. This is the single most useful thing to test yourself in one evening, because the answer for your particular child is not in anyone's marketing — ours included, which is why our own row states no band either.

The parent column is where the category is weakest. Four of the seven competing products say nothing at all about what a parent can see. Khanmigo offers chat history plus moderation alerts and Askie offers controls over what the child can do — two genuinely different things sold under the same word. Photomath is the instructive one: it names parents as people it helps and then describes nothing for them to look at, which is the most common shape in this category and the reason the distinction has its own page.

AI apps a parent can actually monitor — alert, transcript, or record

How to test any of these in one evening

1. 1. Give it the hardest thing your child is actually working on Not a demo prompt — the real worksheet, the real story, the thing that caused the argument on Tuesday. A product that handles a tidy sample and falls apart on your child's actual homework has told you what you need to know in four minutes.
2. 2. Watch whether it withholds the answer Ask it something your child would ask. If it produces the finished answer immediately and every time, you have a vending machine, and no amount of content library changes that. The products worth paying for make the child do the next step.
3. 3. Check the reading level against your own child Have your child read one of its explanations aloud. If they stumble on the vocabulary, the product is pitched above them and you will be translating every session. This is the test the stated age range cannot do for you.
4. 4. Open the parent side before you pay, not after Find the parent view and look at what is actually in it. “Dashboard” on a marketing page can mean a transcript, a list of toggles, or a genuine record of what was learned. Three different products; one word.
5. 5. Ask your child to explain one thing they made with it Thirty seconds, and it is the whole quality gate. A child who can explain it learned something. A child who cannot collected an output. Run this on the trial, before the card renews, and the decision makes itself.

If you only want one recommendation, pick by what you are trying to fix

• You want your child making things, not consuming — Scratch first, and it costs nothing. It is the best on-ramp to creative computing there is, and if your child takes to it you have learned something important about them for free. Tynker and Code.org sit in the same lane with more structure and a syllabus.
• You want homework help that does not just answer — This is the narrow one. Khanmigo is the cheapest serious option and Khan Academy's content behind it is genuinely good. Photomath is the counter-example worth naming because your child has probably already met it: it promises step-by-step explanations, and on its own free card the “how” and “why” tips are struck out — so what $0 buys is the steps to the answer. Whichever you try, run step 2 above on the exact subject your child struggles with before you commit.
• You want to know what actually happened — This is the question that made us build Luchi, so read this with that bias stated. Of the products above, Khanmigo gives you chat history and alerts, Askie gives you switches, and the rest give you nothing; a record of what your child understood and whether it stuck is a different thing again, and it is what we are built around.
• You want a human in the loop — None of the AI products replace that, and Create & Learn is not trying to — it is live small-group classes with a teacher, which is a different purchase with a different price shape. If your child needs accountability more than they need a tool, start there.

Common questions

What is the best AI app for a 9-year-old?

There is no single answer, and any list that gives you one without asking about your child is guessing. The useful split at nine is by goal: creative making (Scratch, free, start here), structured coding (Tynker or Code.org), homework support (Khanmigo at $4/month as of 9 October 2026, or Photomath if the subject is specifically maths), or guided creation with a parent record (ours). Run the one-evening test above on two of them rather than reading a fifth list.

Are there free AI apps for kids that are actually good?

Yes, and they are among the best things in the category. Scratch says plainly on its own site that it “is free for everyone”, and Code.org — now branding itself CodeAI — states that “every CodeAI course, tool, and activity is free to use”. Neither is an AI chatbot; both are creative and computational tools, which for a primary-schooler is frequently the better purchase anyway.

What does a US family actually pay for this kind of app?

Most per-child subscriptions in this category sit between $8 and $30 per month per child, with annual plans discounting that, and the two genuinely free options above sit at zero. Khanmigo is the notable outlier at $4 a month for up to ten children, and Photomath states $9.99 a month or $69.99 a year. Luchi Family is $99 a month, or $49 a month for founding families, covering every kid in the family, and your first 7 days are $5 — cancel before it renews. We are at the top of that range and it is fair to ask what the difference buys; the answer is the parent record, and if that is not what you need, one of the cheaper options above probably is.

How do I know these prices are still right?

You do not, and neither do we after the capture date. Every figure on this page was read off the product's own site on 9 October 2026 and is linked at the foot of the page. Pricing in this category changes several times a year, so treat the table as a shortlist to go and check rather than as a current price list.

Why is Luchi Family in its own comparison table?

Because leaving it out would be the dishonest version. We make one of these products and we are going to be biased, so the fix is to put our row next to everyone else's with the same five columns and the same sourcing rule, say plainly that we are the most expensive thing on the list, and name the cases where a cheaper product is the better buy. All three of those are above — and the same rule is why our own age cell says no band rather than printing a tidy one.

Where these facts come from

Every competitor claim on this page was read off that product's own live site on 9 October 2026 and checked back against the raw page text, not against a summary of it. The sources, in table order: kidsai.app for Askie; photomath.com for Photomath's description and prices, photomath.com/about for the Google and parent sentences, and photomath.com/privacy for its under-13 statement; khanmigo.ai/parents for Khanmigo's price, age statement and parent features; create-learn.us and create-learn.us/about for Create & Learn; tynker.com/pricing for Tynker; scratchfoundation.org for Scratch; and code.org/about for Code.org / CodeAI.

Three honest limitations. First, several of these sites render their pricing in the browser rather than in the page, so a plain fetch cannot see it — that is why Askie's price cell says the price was not on the page we captured rather than giving a figure. Second, a first pass of this research produced two plausible-sounding facts about Create & Learn that a direct search of its own page could not find. Both were dropped rather than printed. If a cell here says “not stated”, that is a finding about the page, not a gap in the research. Third, “no parent view” for Photomath is an absence, so here is the bound on it: the only pages its own site links are the home page, About, the article index, Install, Privacy and Terms, and none of them describes a parent account or a parent dashboard.

We also deliberately do not repeat any product's claims about its own regulatory compliance, or any product's claims about a competitor. Those appear on several of these sites. Verifying them is not something a comparison table can do, and printing them unverified is how a comparison becomes an advertisement for whoever wrote it.

The column most of these products leave blank

Luchi is built around the part the rest of the category skips: your child builds something real, explains it back in their own words, and you get a note written from what actually happened.

Get on the list

How Luchi teaches

Part of the series: Should my kid learn coding or AI?

Related guides in this series

• Coding vs AI for kids: which first? — Not a versus. Why the two are one skill for a school-age kid — and the order that actually makes sense when you sit your child down.
• Will AI replace coding? — AI changes what coding is, not whether your kid should learn it. The durable skill is telling a computer what you want — and how to check it.
• Is coding still worth learning for kids? — Yes — but the skill shifted from syntax to thinking and directing. Exactly what a kid practices when they build with AI.
• What age should kids start with AI? — A practical, age-by-age answer for 5–12s — when AI helps, when to wait, and what “starting” should actually look like.